Decision Making in Emergency Critical Care

SECTION 4 - Pulmonary Critical Care

Respiratory Failure and Mechanical Ventilation

Jon-Emile S. Kenny and Stephen Ruoss


Acute respiratory failure (ARF) is a life-threatening process that requires rapid identification and treatment by the emergency physician.1 The causes of ARF are myriad, and its clinical presentation ranges from somnolence and hypopnea to profound tachypnea, tachycardia, and agitation. Appropriate diagnosis and treatment of ARF flow from a sound understanding of pulmonary pathophysiology.

Normal respiratory activity has been described as a functional chain, beginning with the central nervous system (CNS) and ending with the thoracic cage, muscles of respiration, pulmonary parenchyma, and the pulmonary vasculature.3,4 A break in any of these links may result in ARF. ARF occurs when the respiratory system is unable to meet the metabolic needs of the tissues.2 Classically, this clinical state is divided into hypoxemic and hypercapnic respiratory failure.


Commonly cited causes of hypoxemia include global alveolar hypoventilation, diffusion impairment, ventilation–perfusion mismatch, and shunting.24 Additionally, hypobaric (high-altitude) conditions, low inspired fraction of oxygen, and desaturated mixed venous blood can also cause hypoxemia.


Alveolar ventilation is the process by which the lungs deliver oxygen to the pulmonary capillaries and rid them of carbon dioxide. The alveolar gas Equation describes the relationship between oxygen and carbon dioxide in the alveolus at a given atmospheric pressure and fraction of inspired oxygen:


(PAO2 = alveolar partial pressure of oxygen, PaCO2 = arterial partial pressure of carbon dioxide, FiO2 = fraction of inspired oxygen)

This Equation does not imply cause and effect; it is merely a description of the relationship between oxygen and carbon dioxide tensions in the alveolus. For instance, when the CNS is depressed as a result of a toxic or anatomical insult, normal alveolar ventilation may be impaired. In this context, the partial pressure of carbon dioxide in the alveolus will rise, and the partial pressure of oxygen in the alveolus will fall; the two processes however, are independent from one another—it is the alveolar hypoventilation that links both abnormalities.

The alveolar–arterial, or A–a, gradient is a calculation that provides a means of assessing how well oxygen is moving from the alveoli to the arterial blood:


An elevated A–a gradient (>10 mm Hg) is seen in patients with diffusion defects, ventilation–perfusion mismatch, or shunting. In the setting of hypoventilation, the difference between the alveolar oxygen concentration (as calculated from the alveolar gas Equation above) and the measured arterial PaO2 is minimal, and therefore the A–a gradient should be normal.

In practice, however, global alveolar hypoventilation is commonly complicated by states that also alter the pulmonary parenchyma (e.g., aspiration, compressive atelectasis). The administration of supplemental oxygen, a common intervention in these patients, can also cause ventilation–perfusion anomalies (See below) and nitrogen atelectasis.2,3 These two factors can raise the A–a gradient, thereby diminishing the clinical utility of this calculation.

Diffusion Impairment

From the alveolus to the erythrocyte, a molecule of oxygen must pass through an alveolar epithelial cell, a small interstitial space, the pulmonary capillary endothelium and then into the erythrocyte. Any disease or disorder that disrupts this passive diffusion process is known as diffusion impairment. Diffusion impairment may contribute to—but rarely drives—hypoxemia. Normally, the concentration of oxygen in the alveolus and an erythrocyte equilibrate one-third of the way through the pulmonary capillary bed (carbon dioxide does so much more rapidly).2,3 This rapid equilibration ensures that even when capillary transit time decreases substantially (e.g., during exercise) there will be no compromise of gas exchange. However, hypoxemia will be amplified when there is an effective decrease in alveolar–capillary surface area (e.g., severe emphysema or severe interstitial lung disease such as pulmonary fibrosis) in conjunction with increased pulmonary blood flow. When this pathologic state exists, the oxygen tension in the erythrocyte is unable to equilibrate with alveolar oxygen tension prior to the erythrocyte's passage through the pulmonary capillary bed.

Ventilation–Perfusion (V/Q) Mismatch

V/Q mismatch is the most common and important mechanism of hypoxemia encountered by the emergency physician.1 Alveolar ventilation (Va) is determined by three variables: (1) respiratory rate (RR), (2) tidal volume (Vt), and (3) dead space fraction (Vd/Vt), as related in the following Equation:


(Va = alveolar ventilation, RR = respiratory rate, Vd = dead space ventilation, Vt = tidal volume)

The product of the RR and Vt is known as the minute ventilation (Mve). If alveolar ventilation to a unit of lung exceeds its blood flow (Q), the Va/Q ratio exceeds 1. Such Va/Q mismatch produces alveolar and capillary gas tensions similar to air (i.e., high oxygen and low carbon dioxide). While one would expect such lung units to be beneficial, they are, in fact, inefficient. Lung units with Va/Q ratios >1 yield a high partial pressure of oxygen in the pulmonary capillary; however, the arterial oxygen content draining from these units does not dramatically increase. Recall that the total oxygen content of blood is primarily determined by hemoglobin concentration and saturation, and that only a small portion of blood oxygen content is composed of dissolved oxygen. Because the oxyhemoglobin dissociation curve (Fig. 8.1) is largely flat once the PaO2 is much >60 mm Hg, large increases in PaO2 do not greatly increase oxygen saturation.

FIGURE 8.1 Oxyhemoglobin dissociation curve

Unlike the oxyhemoglobin dissociation curve, the carbon dioxide dissociation curve is linear. Therefore, decrements in the partial pressure of carbon dioxide in high Va/Q lung units result in decreased carbon dioxide concentration, and therefore content, in pulmonary capillary blood. This is why normal, or low, carbon dioxide levels frequently accompany hypoxemia; the small increase in pH caused by carbon dioxide retention is a potent stimulus to augment alveolar ventilation and lower the PaCO2.

Dead space is an area of the lung that is ventilated but not perfused. There is a normal amount of physiologic dead space in all lungs as the conducting airways do not participate in gas exchange. However, alveoli that are not perfused are considered pathologic dead space. These two types of dead space are referred to as anatomical and physiologic dead space, respectively. Dead space (both anatomical and physiologic) is, by mathematical definition, a high Va/Q unit. However, rather than being a function of increased ventilation, dead space is the result of absent pulmonary blood flow (i.e., a Va/Q ratio of infinity). These portions of the lung behave differently from lung units with Va/Q ratios >1 (i.e., high Va/Q ratios). Examination of Equation 8.3 (above) reveals that dead space and alveolar ventilation (Va) are inversely proportional. Therefore, increased dead space results in a functionally low Va/Q physiology. As discussed below, the consequence of low Va/Q physiology is both hypoxemia and hypercapnia.

When alveolar ventilation (Va) falls in relation to perfusion (i.e., a low Va/Q), alveolar gas approaches the composition of mixed venous blood, resulting in a low partial pressure of oxygen and a high partial pressure of carbon dioxide. Consequently, these lung units result in low oxygen- and high carbon dioxide–containing blood. Because blood flow from low Va/Q lung units is by definition proportionately large relative to blood flow from normal or high Va/Q lung units, these lung units contribute disproportionately to the composite arterial blood gas values.


Shunt is the most severe form of low Va/Q physiology (i.e., a Va/Q ratio of zero); shunt occurs when mixed venous blood passes into the left heart without participating in gas exchange. Shunt may have cardiac (e.g., patent foramen ovale) and/or pulmonary (e.g., severe pneumonia and acute respiratory distress syndrome [ARDS]) etiologies. The result of shunt is that mixed venous blood returns to the left heart without being exposed to alveolar gas. The more shunt a lung has, the more mixed venous blood oxygen content will contribute to arterial oxygen content. While it is commonly taught that shunt physiology does not respond to supplemental oxygen, this is only partly true. Oxygen-refractory hypoxemia due to shunt only begins to occur when the shunt fraction of the lung approaches 40% to 50%.3Notably, acute lung injury and ARDS typically occur when the shunt fraction approaches 20% to 30%.

Arterial oxygen saturation is determined by the mixed venous oxygen saturation—that is, the blood entering the lungs from the right ventricle—and the degree to which the lungs can match ventilation with perfusion. If the blood returning to the lungs is disproportionately desaturated—as a result of increased oxygen consumption in the tissues, low cardiac output, or low hemoglobin levels—then the effects of Va/Q mismatch, and in particular shunt physiology, will be magnified.


Hypercapnic respiratory failure is a result of impaired alveolar ventilation (Va). As described in Equation 8.3, alveolar ventilation is determined by three variables: (1) RR, (2) tidal volume, and (3) dead space fraction. It follows that hypercapnia can result from impaired central drive to respiration, neuromuscular weakness, chest wall deformities, lung disease that increases the resistive or elastic load on the lungs, and increased dead space. Many of the aforementioned disease states are also associated with Va/Q mismatch.

While it is intuitive how CNS depression and neuromuscular weakness lead to hypoventilation and hypercapnia, it is less obvious how diseases with increased pulmonary elastic and resistive loads like chronic obstructive pulmonary disease (COPD), asthma, and chronic heart failure (CHF) result in hypercapnia, as patients with these issues typically present with dramatically increased RRs. The important physiologic anomaly in these disease states is rapid shallow breathing. The shallow tidal volume (Vt) serves, despite the tachypnea, to both decrease minute ventilation (Mve) and increase the dead space fraction of ventilation. Because anatomical dead space (Vd) stays relatively constant, a precipitous drop in Vt acts to increase dead space fraction (Vd/Vt). As described above, true dead space ventilation prevents the lung from removing carbon dioxide, and while dead space is technically a high Va/Q ratio, its physiology mirrors that of low Va/Q units with hypoxemia and hypercapnia as a consequence.

The differential diagnosis of arterial hypercapnia should also include increased carbon dioxide production. Arterial carbon dioxide content is directly proportional to the tissue production of carbon dioxide and indirectly proportional to alveolar ventilation (Va).


(VCO2 = carbon dioxide production)

While increased production rarely is the sole cause of hypercapnia, it can be an important contributor, especially when work of breathing is high. With extremis, the muscles of respiration can increase total body carbon dioxide production by fourfold.5 Fever also increases CO2 production by approximately 10% per degree celsius.6


The traditional classification of respiratory failure as either hypoxemic or hypercapneic does provide some indication of underlying etiology of failure, and can help direct initial ventilator settings. Given the overlap between these etiologies discussed above, and the fact that Va/Q mismatch is by far the most common cause of ARF, a more physiologically intuitive and clinically useful approach to ARF may be to overlay the common insults to alveolar ventilation and perfusion onto the framework of the respiratory chain. To this end, a simplified scheme (Table 8.1) thus divides ARF into neuromuscular abnormalities and parenchymal abnormalities which include airway injury or dysfunction, alveolar injury, and pulmonary vascular injury.2

TABLE 8.1 Causes of Acute Respiratory Failure


Mastering the basic nomenclature of mechanical ventilation is challenging, and is complicated by inconsistent naming among manufacturers and by novel ventilation modes available with newer devices. This section outlines the basics of mechanical ventilation. Ventilation strategies for specific disease entities are elaborated upon in following chapters.

Modes of Invasive Mechanical Ventilation

Invasive ventilation entails the application of positive pressure via an endotracheal tube. The breath type delivered to the patient defines a mode of ventilation. Breath types, in turn, are defined by three variables: what triggers (initiates), limits (maintains), and cycles (terminates) a breath. While there are three variables, it is typically how a breath is cycled that categorizes the ventilation mode. Volume-cycled breaths are terminated when a preset volume has been achieved. Pressure-cycled breaths are terminated when a preset time has been reached (Fig. 8.2). While the later are technically time-cycled breaths, the common clinical parlance of ‘pressure-cycled’ is used here.

FIGURE 8.2 Volume-cycled versus pressure-cycled breaths

The choice between using volume-cycled and pressure-cycled modes of ventilation depends mostly on what the clinician desires to control. When a patient needs a guaranteed Mve (e.g., a patient with severe acid–base disturbances), it is prudent to choose a volume-cycled mode of ventilation because Mve is controlled directly. However, when airway pressure needs to be strictly managed (e.g., in a patient at risk of ventilator-induced lung injury and/or high airway pressures) then a pressure-cycled mode of mechanical ventilation should be instituted.

Importantly, when a clinician initiates and monitors ventilation that is volume cycled, there will be varying peak and plateau pressures depending on airway resistance and thoracic compliance, respectively. Conversely, Vt—and therefore Mve—will vary when pressure is the predetermined variable. A more detailed discussion of thoracic compliance and the relationship between airway pressure and Vt is presented below.

Volume-Cycled Ventilation

The two most common modes of volume-cycled ventilation are volume assist–control ventilation (AC; also known simply as assist–control), and synchronized intermittent mandatory ventilation (SIMV). AC is defined by delivery of a set Vt for each breath, regardless of whether the breath is initiated by the ventilator or the patient. In AC mode, the patient receives, at a minimum, the ventilation rate and the tidal volume set on the ventilator. If a patient has an intrinsic respiratory drive to breathe at a rate faster than the one set on the mechanical ventilator in AC mode, the patient will still receive the full machine-delivered tidal volume for every breath initiated. Thus, AC ventilation mode involves assured delivery of a set and consistent Vt for any net RR, regardless of whether the patient or the machine is initiating the breath.

In contrast to AC, the central ventilation feature of SIMV is the delivery of the preset Vt to the patient only at the rate set on the ventilator. Breaths initiated by the patient over the set machine ventilation rate are not assisted with a preset Vt, but instead follow a Vt that is generated independently by the patient. Preset inspiratory pressure assistance—or pressure support—is an added feature that helps the patient to achieve a physiologically reasonable Vt in the absence of a set Vt for breaths initiated by the patient above and beyond the set RR.

Both AC and SIMV guarantee a minimum Mve, because the clinician directly sets both RR and Vt. An important operational difference between AC and SIMV is that AC will allow full Vt support for a patient breathing over the preset ventilation rate; in SIMV, if the preset ventilator rate is not sufficient for the patient's ventilator requirements, the patient will not receive an adequate Mve. This is particularly important when initiating ventilation, as the true physiologic needs of the patient may not yet be fully understood. Thus, in initial ventilator mode setup, including in an emergency department (ED), there may be an advantage in choosing AC. For the initial treatment of sedated and paralyzed ED patients, there will be no difference in the achieved minute ventilation.

Pressure-Cycled Ventilation

In contrast to SIMV and AC, pressure-cycled modes of ventilation use airway pressure as the independent (clinician-controlled) variable. In these modes, therefore, Vt (and consequently Mve) becomes the dependent variable and is a function of the preset pressure, airway resistance, and thoracic compliance. Pressure control ventilation (PCV) is analogous to volume assist–control, in that the ventilator can deliver both assisted (patient-triggered) and controlled (machine-triggered) breaths; however, they are pressure (not volume) cycled. As noted above, in PCV, both assisted and controlled breaths are time-cycled. Therefore, in these modes, direct control can be maintained over inspiratory and expiratory time (i.e., the I:E ratio). PCV is typically utilized when the clinician desires direct control over airway pressure at the expense of guaranteed volume.

Pressure Support Ventilation

In this mode, the ventilator provides a preset pressure throughout spontaneous patient inspiration, leaving the patient to control inspiratory and expiratory times as well as achieved Vt. Given the dependence of pressure support ventilation (PSV) on patient cooperation and effort, respiratory support provided by this mode can be altered substantially by patient sedation, respiratory muscular weakness, or clinical features such as pain or agitation. PSV is commonly used in the ICU prior to extubation (i.e., as a “weaning” mode), and so is unlikely to be encountered in the ED.

Noninvasive Positive Pressure Ventilation

Noninvasive positive pressure ventilation (NIPPV) entails the application of positive pressure to the patient via a tight-fitting mask over the nose or mouth and nose. Essentially, there are two modes of NIPPV—bilevel positive airway pressure (BiPAP) and continuous positive airway pressure (CPAP). As the name suggests, BiPAP requires that the clinician set two pressure variables: the inspiratory positive airway pressure (IPAP) and the expiratory positive airway pressure (EPAP). The change in pressure in BiPAP—referred to as the “delta”—is the IPAP minus the EPAP, because both pressures are referenced to zero (atmospheric) pressure. The difference between inspiratory and expiratory pressure is the driving pressure to support alveolar ventilation and CO2 clearance. Note that the EPAP is equivalent to the positive end-expiratory pressure, or PEEP, value used with invasive modes of ventilation, and is defined as the set pressure above atmospheric pressure that is delivered throughout expiration. BiPAP can be thought of as the noninvasive equivalent to PSV. One important difference between PSV and BiPAP is the terminology used to define the pressure parameters. In PSV, the inspiratory pressure is delivered above a baseline PEEP. In BiPAP, the IPAP is always delivered above atmospheric pressure, not above EPAP.

CPAP, by contrast, requires only one pressure preset. A single pressure is delivered throughout the respiratory cycle; that is, the IPAP and EPAP are the same. CPAP mainly benefits oxygenation by stenting open collapsed airways, thereby reducing the number of low Va/Q lung units. Because BiPAP has preset inspiratory and expiratory pressures, BiPAP increases the Vt and therefore aids in ventilation (CO2 elimination) as well as oxygenation.


Two important physiologic aspects of mechanical ventilator support are airway pressures and PEEP. The peak airway pressure (also known as the peak inspiratory pressure or PIP) measured by the ventilator is the pressure required to overcome both the thoracic (lung and chest wall) compliance and the airway resistance (Equation 8.5). Compliance describes the change in volume with respect to a change in pressure in a deformable object. For example, a poorly compliant thorax has a small change in volume for a large change in pressure. Airway resistance, analogous to resistance in blood vessels, describes the mechanical factors that limit airflow. There are approximately 23 generations, or branching points, of airways from the trachea to the alveoli, and the resistance to airflow is highly dependent upon the cumulative cross-sectional area of each generation in parallel. While the trachea has a larger diameter than a single terminal bronchiole, the trachea has a much smaller diameter than the entire cross-sectional diameter of all the terminal bronchioles in parallel. Hence, the trachea contributes more to airway resistance than do the terminal bronchioles in the healthy lung.


(PIP = peak inspiratory pressure, Vt = tidal volume, Ct = thoracic compliance, Raw = airway resistance, Q = airflow)

Note that Equation 8.5 divides PIP into two components—a static component that is determined by the Vt and the compliance of the thorax, and a dynamic component that is determined by the flow of gas and the composite resistance of the lungs' airways. Consequently, an increase in airway resistance or a decrease in thoracic compliance will increase peak airway pressure. Distinguishing between the two requires a cessation of airflow at end-inspiration. In the absence of airflow (i.e., Q is zero), the resistive component is removed from the Equation, and the pressure that remains is related only to the thoracic compliance; this pressure is referred to as the plateau pressure.

When evaluating a patient with high PIPs, a large pressure drop between the peak and plateau pressure suggests an excess of airway resistance. Conversely, if there is little difference between the peak and plateau pressures, there is likely a poorly compliant lung or chest wall (Fig. 8.3). Normal peak and plateau values are approximately 20 and 10 cm H2O, respectively. An inspiratory hold maneuver is best carried out in a volume-cycled mode of ventilation rather than a pressure-cycled one. This is because in volume-cycled ventilation, pressure is the dependent variable.

FIGURE 8.3 Peak and plateau pressures

PEEP may be applied to a patient supported with either invasive or noninvasive ventilation. It is often applied to some small degree under the guise of replacing the “physiologic PEEP,” which is reportedly lost when the endotracheal tube separates the vocal cords.7 While there is little evidence that physiologic PEEP exists, PEEP can be applied therapeutically to aid in oxygenation.8 Typical PEEP levels range between 5 and 15 cm H2O. PEEP promotes oxygenation by preventing alveolar and small airway closure at end-expiration. As previously noted, multiple mechanisms may result in low Va/Q lung units. For example, increased airway resistance due to inflammation, secretions and airway edema; physical compression secondary to habitus; and excessively compliant airways may all lead to low alveolar ventilation relative to perfusion. The judicious application of PEEP preserves patent airways and maintains the lung on a mechanically favorable portion of its compliance curve.

While PEEP has beneficial applications, it can be detrimental to both the heart and lungs. PEEP has a profound effect on mean airway pressure and can exert a multitude of effects on the right heart, pulmonary vasculature, and left heart. Specifically, excessive PEEP can impair venous return, which in turn has the potential to diminish cardiac output and compromise oxygen delivery to the tissues.9Excessive PEEP can also lead to alveolar rupture.10

The time constant of a lung unit describes the length of time required for inflation and deflation of a ventilated portion of the lung. The time constant is directly proportional to the resistance and compliance of the lung unit. Hence, if the resistance or the compliance of a lung unit increases, the time it takes to deflate increases. This is particularly important in patients with emphysema, as both resistance and compliance may be dramatically elevated. If lung units fail to deflate before a subsequent breath is taken (or delivered by a ventilator), retained volume—and consequently pressure—can build within portions of the lung. This phenomenon is called auto-PEEP or intrinsic PEEP. The risk of auto-PEEP becomes greatest in patients with airway obstruction and tachypnea. In order to detect auto-PEEP, the clinician must first anticipate its existence. On the ventilator, the presence of an expiratory flow curve that does not reach zero flow prior to a subsequent breath is suggestive of auto-PEEP (Fig.8.4). Since the pressure at end-expiration is the sum total of both extrinsic (machine delivered) and intrinsic PEEP, an end-expiratory breath hold maneuver is another way to reveal pressure present in the airway. Just as with extrinsic PEEP, excessive auto-PEEP can diminish venous return to the right heart and negatively impact cardiac output. Treatment of auto-PEEP that results in hemodynamic compromise includes briefly removing the patient from the ventilator to allow for lung and chest decompression. For ongoing correction of intrinsic PEEP, sedating the patient and/or lowering the set RR and decreasing the inspiratory time (and thus prolonging the expiratory time at any given ventilation rate) will all help avoid or reverse intrinsic PEEP.



Indications for mechanical ventilation can be difficult to define for all clinical conditions. Often, it is a gestalt decision based upon the clinical status of the patient in combination with objective measures such as pulse oximetry and arterial blood gas sampling (Table 8.2). When neurologic insult impairs global alveolar ventilation, the resultant hypoxemia and hypercapnia are usually easily managed with invasive mechanical ventilation. A volume-cycled mode such as AC or SIMV is preferred, because the clinician can directly control Mve while monitoring airway pressure. In the absence of severe pulmonary parenchymal abnormalities, minimal PEEP is typically required. NIPPV should be avoided, as the unconscious or obtunded patient is at high risk for gastric distention and aspiration without definitive control of the airway.

TABLE 8.2 Indications for Mechanical Ventilation

When ventilating a patient with severe obstructive lung disease, the clinician must be wary of auto-PEEP and dynamic hyperinflation. It can be difficult to balance the need for increased Mve for the treatment of these patients' hypercapnia with the potential for producing auto-PEEP. Usually a volume-cycled mode (A/C or SIMV) is chosen for control of Mve, coupled with short inspiratory times to allow for adequate expiration. Often, sedation is required to synchronize the patient with the ventilator. Permissive hypercapnia (allowance for a modest supranormal increase in PaCO2) may be needed to allow for complete emptying of the lungs at a lower RR. Externally applied PEEP can be useful in emphysema, a condition in which increased airway compliance makes the alveoli susceptible to collapse during positive pressure ventilation (See Chapter 9). Furthermore, when auto-PEEP exists, extrinsic PEEP may decrease the pressure gradient required to trigger the ventilator. While adding extrinsic PEEP to aid a patient with excessive intrinsic PEEP may seem counterintuitive, the rationale is that ventilators use small deflections in airway pressure as the signal to begin a subsequent breath (pressure triggered). Ventilators may also use changes in flow and time as other variables to trigger breaths. The drop in pressure that must be achieved to trigger a breath is referenced to the preset (extrinsic) PEEP. If there is superimposed intrinsic PEEP, then the patient must create a pressure drop that includes all of the intrinsic PEEP plus the trigger value below the extrinsic PEEP. Increasing the extrinsic PEEP will narrow this pressure differential and ease breath triggering.11

When managing alveolar and/or interstitial edema, it is recommended to adopt a low-lung-volume ventilation strategy, especially if ARDS is the presumed cause of ARF.2 In this paradigm, plateau pressures of <30 cm H2O are desirable. If a volume-cycled mode of ventilation is chosen, peak and plateau pressures should be closely monitored. This hazard can be avoided by placing the patient on a pressure-cycled mode (e.g., PCV), which allows the clinician to select the delivered pressure. The trade-off of selecting a pressure-cycled mode, however, is that close observation of Vt and Mve (the dependent variables) is required. As ventilators become more sophisticated, modes of ventilation that deliver safe levels of pressure to obtain a preset volume (e.g., pressure-regulated volume control) are being adopted. However, the only mode of ventilation known to improve mortality in ARDS is volume control, as this was used in the ARDSNet trial.2

The use of NIPPV for COPD exacerbations has been demonstrated to decrease mortality and intubation rates, and improved long-term outcomes.13 Evidence for the application of NIPPV in acute asthma is less robust.14 While there is evidence to support the use of NIPPV in cardiogenic pulmonary edema to improve dyspnea, gas exchange, and perhaps prevent intubation,15,16 the data for ARDS are less definitive.17,18


Acute respiratory failure is a life-threatening process that requires rapid identification and treatment. The decision to initiate mechanical ventilation is always best made by the emergency physician at bedside, however a nuanced understanding of the different etiologies of respiratory failure and optimal corresponding modes of ventilatory support described in this chapter can help ensure patient safety and improve outcomes.


CI, confidence interval; RR, relative risk.


1.Stefan MS, Shieh MS, Pekow PS, et al. Epidemiology and outcomes of acute respiratory failure in the United States, 2001 to 2009: a national survey. J Hosp Med. 2013;8:76–82.

2.Greene KE, Peters JI. Pathophysiology of acute respiratory failure. Clin Chest Med. 1994;15:1–12.

3.D'Alonzo GE, Dantzger DR. Mechanisms of abnormal gas exchange. Med Clin North Am. 1983;67:557–571.

4.MacSweeney RM, McAuley DF, Matthay MA. Acute lung failure. Semin Respir Crit Care Med. 2011;32:607–625.

5.Roussos C, Koutsoukou A. Respiratory failure. Eur Respir J. 2003;22:3S–14S.

6.Manthous CA, Hall JB, et al. Effect of cooling on oxygen consumption in febrile critically ill patients. Am J Respir Crit Care Med. 1995;151:10–14.

7.Fernández-Mondéjar E. Prophylactic positive end-expiratory pressure: are good intentions enough? Crit Care. 2003;7:191.

8.Rimensberger PC, Cox PN, et al. The open lung during small tidal volume ventilation: concepts of recruitment and “optimal” positive end-expiratory pressure. Crit Care Med. 1999;27(9):1946–1952.

9.Lueke T, Pelosi P. Clinical review: positive end-expiratory pressure and cardiac output. Crit Care. 2005;9(6):607–621.

10.Dos Santos CC, Slutsky AS. Mechanisms of ventilator-induced lung injury: a perspective. J Appl Physiol. 2000;89:1645–1655.

11.MacIntyre NR, et al. Applied PEEP during pressure support reduces the inspiratory threshold load of intrinsic PEEP. Chest. 1997;111(1):188–193.

12.The Acute Respiratory Distress Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med. 2000;342(18):1301–1308.

13.Ram FS, Picot J, et al. Non-invasive positive pressure ventilation for treatment of respiratory failure due to exacerbations of chronic obstructive pulmonary disease. Cochrane Database Syst Rev. 2004;(3):CD004104.

14.Soroksky A, et al. A pilot prospective, randomized, placebo-controlled trial of bilevel positive airway pressure in acute asthmatic attack. Chest. 2003;123(4):1018.

15.Gray A, Goodacre S, et al. 3CPO Trialists. Noninvasive ventilation in acute cardiogenic pulmonary edema. N Engl J Med. 2008;359(2):142–151.

16.Weng CL, Zhao YT, et al. Meta-analysis: noninvasive ventilation in acute cardiogenic pulmonary edema. Ann Intern Med. 2010;152(9):590–600.

17.Agarwal R, Aggarwal AN, Gupta D. Role of noninvasive ventilation in acute lung injury/acute respiratory distress syndrome: a proportion meta-analysis. Respir Care. 2010;55(12):1653–1660.

18.Zhan Q, et al. Early use of noninvasive positive pressure ventilation for acute lung injury; a multicenter randomized controlled trial. Crit Care Med. 2012;40(2):455.